Scientists from the Institute of Petroleum Geology and Geophysics of the Siberian Branch of the Russian Academy of Sciences (INGG SB RAS) have developed a neural network algorithm that dramatically accelerates the search for common mineral resources. As TASS was informed by the institute's press service, the methodology has already been tested in the southeast of the Yamal-Nenets Autonomous Okrug.

The algorithm processes archival 3D seismic survey data, focusing on the upper part of the geological section with a thickness of 30 to 100 meters – this is where sands, clays, gravel, and peat are found. For large infrastructure projects in the Yamal-Nenets Autonomous Okrug, the volume of sand raw materials can exceed 1 million cubic meters, and the distance of quarries reaches 300 km, making transportation costs critical.

The development solves the problem of expensive logistics: promising areas are identified directly within the licensed territory of hydrocarbon fields. The neural network is resistant to noise and does not require retraining for each new area, which favorably distinguishes it from traditional approaches. In essence, old seismic arrays are transformed into a ready-made map of building materials, reducing environmental damage and the cost of developing northern fields.